Dice Question Streamline Icon: https://streamlinehq.com

Local Interaction Rules for Complex Shepherding Tasks

Engineer and validate local interaction rules among controller–controller and controller–target agents that realize more complex shepherding tasks within the harnessing-complex-systems-for-control framework.

Information Square Streamline Icon: https://streamlinehq.com

Background

After presenting shepherding as a key instance of harnessing one complex system to control another, the chapter notes that existing interaction rules often target basic objectives (e.g., gathering). Extending these rules to more complex tasks requires principled design that maps local interactions to desired global behaviors.

The authors explicitly flag the need to design richer local interaction mechanisms to broaden the class of achievable shepherding objectives, moving beyond simple or canonical tasks.

References

Several open challenges remain in harnessing complex systems for control. These include developing continuum models to describe emergent shepherding behavior, engineering local interaction rules for more complex tasks, addressing scenarios with actively escaping targets, and extending the framework to three-dimensional spaces and other geometries.

Controlling Complex Systems (2504.07579 - Coraggio et al., 10 Apr 2025) in Subsection "Harnessing Complex Systems for Control"